U.S. patent application number 15/400610 was filed with the patent office on 2017-04-27 for image capturing apparatus, image capture system, image processing method, information processing apparatus, and computer-readable storage medium.
The applicant listed for this patent is Daisuke BESSHO, Toru HARADA, Nozomi IMAE, Yoshiaki IRINO, Yoichi ITO, Kensuke MASUDA, Hiroyuki SATOH, Satoshi SAWAGUCHI, Makoto SHOHARA, HIROKAZU TAKENAKA, Tomonori TANAKA, Noriyuki TERAO, Hideaki YAMAMOTO. Invention is credited to Daisuke BESSHO, Toru HARADA, Nozomi IMAE, Yoshiaki IRINO, Yoichi ITO, Kensuke MASUDA, Hiroyuki SATOH, Satoshi SAWAGUCHI, Makoto SHOHARA, HIROKAZU TAKENAKA, Tomonori TANAKA, Noriyuki TERAO, Hideaki YAMAMOTO.
Application Number | 20170116704 15/400610 |
Document ID | / |
Family ID | 49116920 |
Filed Date | 2017-04-27 |
United States Patent
Application |
20170116704 |
Kind Code |
A1 |
TAKENAKA; HIROKAZU ; et
al. |
April 27, 2017 |
IMAGE CAPTURING APPARATUS, IMAGE CAPTURE SYSTEM, IMAGE PROCESSING
METHOD, INFORMATION PROCESSING APPARATUS, AND COMPUTER-READABLE
STORAGE MEDIUM
Abstract
An image capturing apparatus includes a tilt detection unit that
detects a tilt in the vertical direction, conversion data used for
transforming plane coordinates into spherical coordinates, a
correction unit that corrects the conversion data according to the
tilt, a plurality of image capturing units, a coordinate
transformation unit that transforms plane coordinates of a
plurality of pixels included in images captured by the image
capturing units into spherical coordinates according to the
conversion data corrected by the correction unit, and a combining
unit that combines the images including the pixels transformed into
spherical coordinates by the coordinate transformation unit.
Inventors: |
TAKENAKA; HIROKAZU;
(Kanagawa, JP) ; IRINO; Yoshiaki; (Kanagawa,
JP) ; TANAKA; Tomonori; (Kanagawa, JP) ; IMAE;
Nozomi; (Kanagawa, JP) ; YAMAMOTO; Hideaki;
(Kanagawa, JP) ; MASUDA; Kensuke; (Kanagawa,
JP) ; ITO; Yoichi; (Tokyo, JP) ; SAWAGUCHI;
Satoshi; (Kanagawa, JP) ; BESSHO; Daisuke;
(Kanagawa, JP) ; SATOH; Hiroyuki; (Kanagawa,
JP) ; SHOHARA; Makoto; (Tokyo, JP) ; TERAO;
Noriyuki; (Miyagi, JP) ; HARADA; Toru;
(Kanagawa, JP) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
TAKENAKA; HIROKAZU
IRINO; Yoshiaki
TANAKA; Tomonori
IMAE; Nozomi
YAMAMOTO; Hideaki
MASUDA; Kensuke
ITO; Yoichi
SAWAGUCHI; Satoshi
BESSHO; Daisuke
SATOH; Hiroyuki
SHOHARA; Makoto
TERAO; Noriyuki
HARADA; Toru |
Kanagawa
Kanagawa
Kanagawa
Kanagawa
Kanagawa
Kanagawa
Tokyo
Kanagawa
Kanagawa
Kanagawa
Tokyo
Miyagi
Kanagawa |
|
JP
JP
JP
JP
JP
JP
JP
JP
JP
JP
JP
JP
JP |
|
|
Family ID: |
49116920 |
Appl. No.: |
15/400610 |
Filed: |
January 6, 2017 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
14384027 |
Sep 9, 2014 |
|
|
|
PCT/JP2013/057338 |
Mar 11, 2013 |
|
|
|
15400610 |
|
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06T 3/0062 20130101;
H04N 5/23238 20130101 |
International
Class: |
G06T 3/00 20060101
G06T003/00; H04N 5/232 20060101 H04N005/232 |
Foreign Application Data
Date |
Code |
Application Number |
Mar 9, 2012 |
JP |
2012-053190 |
Nov 30, 2012 |
JP |
2012-263542 |
Claims
1. (canceled)
2. An image capturing apparatus, comprising: at least two lenses
with different image capturing directions; a sensor to detect a
tilt of the image capturing apparatus; and circuitry configured to
transform an image captured through the at least two lenses, based
on the tilt which has been detected by the sensor.
3. The image capturing apparatus according to claim 2, wherein the
circuitry configured to transform transforms plane coordinates of
the images captured by the image capturing apparatus into spherical
coordinates, based on the tilt which has been detected.
4. The image capturing apparatus according to claim 2, wherein the
two lenses face in opposite directions.
5. The image capturing apparatus according to claim 2, further
comprising two image capturing elements, corresponding to the at
least two lenses, that face in opposite directions.
6. The image capturing apparatus according to claim 2, further
comprising: circuitry configured to correct conversion data used
for transforming plane coordinates into spherical coordinates,
based on the tilt which has been detected.
7. The image capturing apparatus according to claim 2, wherein:
each of the at least two lenses comprises a wide-angle lens.
8. The image capturing apparatus according to claim 2, wherein:
each of the at least two lenses comprises a lens with an angle of
view equal to or larger than 180 degrees.
9. An information processing apparatus, comprising: circuitry
configured to transform an image captured through at least two
lenses with different image capturing directions of an image
capturing apparatus, based on a tilt of the image capturing
apparatus, the tilt having been detected by a sensor.
10. The information processing apparatus according to claim 9,
wherein: each of the at least two lenses comprises a wide-angle
lens.
11. The information processing apparatus according to claim 9,
wherein: each of the at least two lenses comprises a lens with an
angle of view equal to or larger than 180 degrees.
12. An image processing method, comprising: transforming an image
captured through at least two lenses with different image capturing
directions of an image capturing apparatus, based on a tilt of the
image capturing apparatus, the tilt being detected by a sensor.
13. The image processing method according to claim 12, wherein:
each of the at least two lenses comprises a wide-angle lens.
14. The image processing method according to claim 12, wherein:
each of the at least two lenses comprises a lens with an angle of
view equal to or larger than 180 degrees.
Description
CROSS REFERENCE TO RELATED APPLICATIONS
[0001] This patent application is a Continuation Application of
U.S. application Ser. No. 14/384,027, filed Sep. 9, 2014, which is
a National Stage Application of International Application No.
PCT/JP13/057338, filed Mar. 11, 2013, which is based on and claims
priority to Japanese Patent Application Nos. 2012-263542, filed
Nov. 30, 2012, and 2012-053190, filed Mar. 9, 2012, the entire
contents of each of the above are hereby incorporated by reference
herein.
TECHNICAL FIELD
[0002] The present invention relates to an image capturing
apparatus, an image capture system, an information processing
apparatus, and a computer-readable storage medium.
BACKGROUND ART
[0003] Conventional image capturing apparatuses that capture images
in all directions or spherically employ a hyperboloidal mirror or a
fisheye lens, for example. These image capturing apparatuses are
specific optical systems that need a non-linear image processing
such as distortion correction or projective transformation so that
the captured image with the apparatus can be visualized to a
user.
[0004] One of known technologies is that a semispherical image is
obtained with a wide-angle capture employing a fisheye lens and
recorded as a distorted circular image. A part of the distorted
circular image is cut out and image processing is performed on the
image with a computer, whereby the distorted image is transformed
into a plane regular image.
[0005] Another known technology is that while performing the image
processing above, when the center position of the distorted
circular image does not correctly correspond to the vertical point
direction, a user specifies a parameter of a tilt angle. The load
while performing distortion correction on the image is therefore
reduced.
[0006] As described above, when images are captured in a state of
the image capturing apparatus being tilted, an omni-directional
spherical image in the wrong vertical direction may be generated.
Some inventions to solve this problem have been already known.
[0007] Japanese Patent Application Laid-open No. 2003-223633 and
Japanese Patent Application Laid-open No. 2006-059202 disclose a
technology of generating an omni-directional spherical image having
the correct vertical direction. This is achieved by adding
rotational transformation according to the tilt of a camera to the
process of non-linear image processing.
[0008] Japanese Patent Application Laid-open No. H11-309137 and
Japanese Patent No. 4175832 disclose a technology of generating an
omni-directional spherical image having the correct vertical
direction at high speed, for the purpose of reducing the time
required for generating the omni-directional spherical image having
the correct vertical direction or reducing the cost on arithmetic
operations. This technology is achieved by adding rotational
transformation according to the tilt of the camera, in addition to
distortion correction or projective transformation, preparing a
conversion table used for non-linear transformation in advance, and
performing batch transformation through the conversion table when
capturing images.
[0009] With the structure of previous spherical image capturing
apparatuses, however, users are required to specify a tilt angle of
the apparatus during capture when combining a plurality of images
captured at once from a plurality of image capturing directions,
when transforming the combined image into a plane regular image,
and when displaying the resulting image. That is to say, there was
a problem in that a necessary parameter for correction could not be
obtained during capture for automatically correcting the image.
[0010] Additionally, with the previously performed method for
capturing images in all directions or spherically through a
conversion table according to the tilt, if the amount or
orientation of the tilt changes, the conversion table needs to be
re-made from the beginning. In this case, processing requires a
relatively long time to generate an omni-directional spherical
image having the correct vertical direction according to any tilt
of the image capturing apparatus.
[0011] Japanese Patent Application Laid-open No. H11-309137 and
Japanese Patent No. 4175832 disclose in particular a technology
generating images using a conversion table that includes rotational
transformation according to the tilt of the camera. The conversion
table is made so as to correspond to the amount of tilt determined
in advance. Accordingly, if the amount or orientation of the tilt
changes so as to differ from the predetermined amount or
orientation, the conversion table needs to be re-made from the
beginning. Also in this case, the processing still requires a
relatively long time to generate an omni-directional spherical
image having the correct vertical direction according to any tilt
of the image capturing apparatus, therefore the problem is not yet
solved.
[0012] Therefore, there is a need to provide an image capturing
apparatus capable of generating an omni-directional spherical image
having the correct vertical direction, according to any tilt of the
image capturing apparatus.
DISCLOSURE OF INVENTION
[0013] It is an object of the present invention to at least
partially solve the problems in the conventional technology.
[0014] According to an embodiment, there is provided an image
capturing apparatus that includes a tilt detection unit configured
to detect a tilt in a vertical direction; conversion data used for
transforming plane coordinates into spherical coordinates; a
correction unit configured to correct the conversion data according
to the tilt; a plurality of image capturing units; a coordinate
transformation unit configured to transform plane coordinates of a
plurality of pixels included in images captured by the image
capturing units into spherical coordinates according to the
conversion data corrected by the correction unit; and a combining
unit configured to combine the images including the pixels
transformed into spherical coordinates by the coordinate
transformation unit.
[0015] The above and other objects, features, advantages and
technical and industrial significance of this invention will be
better understood by reading the following detailed description of
presently preferred embodiments of the invention, when considered
in connection with the accompanying drawings.
BRIEF DESCRIPTION OF DRAWINGS
[0016] FIG. 1 is a schematic block diagram for explaining the
overall structure of an image capturing apparatus according to an
embodiment of the present invention.
[0017] FIG. 2 is an external side view of an omni-directional
spherical image capturing apparatus according to the
embodiment.
[0018] FIG. 3 is a flowchart for explaining operations of the image
capturing apparatus according to the embodiment.
[0019] FIG. 4 illustrates a projection relation of a fisheye lens
employed in the image capturing apparatus according to the
embodiment.
[0020] FIG. 5 illustrates a format for the omni-directional
spherical image captured by the image capturing apparatus according
to the embodiment.
[0021] FIG. 6 illustrates the outline of generation processing of
the omni-directional spherical image captured by the image
capturing apparatus according to the embodiment.
[0022] FIG. 7 illustrates the conversion table for the
omni-directional spherical image captured by the image capturing
apparatus according to the embodiment.
[0023] FIG. 8 is a flowchart for explaining operations of
correction processing through the conversion table for the
omni-directional spherical image captured by the image capturing
apparatus according to the embodiment.
[0024] FIG. 9 is a schematic diagram for explaining a tilt of the
image capturing apparatus according to the embodiment.
[0025] FIG. 10 illustrates an example of calculation of vertical
correction on the omni-directional spherical image captured by the
image capturing apparatus according to the embodiment.
[0026] FIG. 11 illustrates another example of calculation of
vertical correction on the omni-directional spherical image
captured by the image capturing apparatus according to the
embodiment.
[0027] FIG. 12 is another flowchart for explaining operations of
correction processing through the conversion table for the
omni-directional spherical image captured by the image capturing
apparatus according to the embodiment.
[0028] FIG. 13 is a schematic diagram for explaining the overall
structure of an image capture system according to another
embodiment of the present invention.
[0029] FIG. 14 is a schematic block diagram for explaining the
overall structure of an electronic circuit serving as a destination
device of the image capture system according to another
embodiment.
[0030] FIG. 15 is a flowchart for explaining operations of the
image capture system according to another embodiment.
[0031] FIG. 16 is a diagram for explaining the structure of a
multi-eye image capturing apparatus that captures images in all
directions spherically at once.
[0032] FIG. 17 is a diagram for explaining an image obtained
through a lens in the multi-eye image capturing apparatus
illustrated in FIG. 16.
[0033] FIG. 18 is a diagram for explaining an image obtained by
performing distortion correction on images illustrated in FIG. 17
and then combining the images with each other.
[0034] FIG. 19 is a diagram for explaining an image captured by the
multi-eye image capturing apparatus illustrated in FIG. 16, with
the image capturing apparatus in a state of being tilted.
[0035] FIG. 20 is a diagram for explaining an image obtained by
performing distortion correction and image combination on the image
illustrated in FIG. 19 without consideration of the tilt.
[0036] FIG. 21 is a schematic diagram for explaining an
acceleration sensor that measures a tilt angle.
BEST MODE(S) FOR CARRYING OUT THE INVENTION
[0037] Exemplary embodiments according to the present invention are
described below with reference to the accompanying drawings. In the
drawings and descriptions below, the same or equivalent components
are designated by the same reference numerals, and overlapped
explanation thereof will be simplified or omitted as appropriate.
In the embodiments, when generating an omni-directional spherical
image, the omni-directional spherical image capturing apparatus
detects the vertical direction and corrects a conversion table used
for image processing according to the vertical direction. In the
embodiments, the omni-directional spherical image is generated
through the corrected conversion table. Accordingly, the conversion
table does not need to be re-made from the beginning, thereby
reducing the processing time. The image capturing apparatus refers
to a stand-alone apparatus, such as a digital camera; an image
capture system refers to a system in which a plurality of
apparatuses such as a digital camera and an information processing
apparatus are utilized separately. In the embodiments, unless
otherwise specified, the image capturing apparatus conceptually
includes an image capture system.
[0038] A multi-eye image capturing apparatus will now be described
that captures images in all directions spherically at once. FIG. 16
is a diagram for explaining the structure of a multi-eye image
capturing apparatus that captures images in all directions
spherically at once. An image capture system is prepared that
employs a plurality of wide-angle lenses as illustrated in FIG. 16
so as to capture images in all directions spherically at once.
[0039] For example, the multi-eye image capturing apparatus
illustrated on the left in FIG. 16 employs a fisheye lens (ultra
wide-angle lens) with an angle of view equal to or larger than 180
degrees. The multi-eye image capturing apparatus requires at least
the two-eye structure (image capture lenses with two different
image capturing directions) to capture images in all directions
spherically at once. If image distortion needs to be reduced as
much as possible, the number of image capture lenses may be
increased by employing, for example, the four-eye structure for the
multi-eye image capturing apparatus as illustrated on the right in
FIG. 16. With this structure, only center parts with high image
quality are selectively used to reduce the image distortion. In
this image capturing apparatus with the four-eye structure, it is
recommended that respective lenses have an angle of view equal to
or larger than 90 degrees, and preferably about 100 degrees.
[0040] The image capturing apparatus with two-eye structure will be
described hereafter, for convenience of description. The basic idea
of the present invention may be applied to the image capturing
apparatus with the four-eye structure illustrated on the right in
FIG. 16, however.
[0041] In the image capturing apparatus, it is assumed that the
central axis of the image capturing apparatus in the vertical
direction corresponds to the vertical axis of the sphere.
Specifically, an image obtained through each of two fisheye lenses
is presented as illustrated in FIG. 17. FIG. 17 is a diagram for
explaining an image obtained through a lens in a two-eye image
capturing apparatus illustrated on the left in FIG. 16. As
described above, it is preferred that images are captured so that
the vertical point and the horizontal line recognized by a user of
the image capturing apparatus correspond to the vertical point and
the horizontal line of the image capturing apparatus.
[0042] The borderline of 180 degrees illustrated in FIG. 17 will be
described. In an image capture system including two lenses (two
image capturing directions), two images are captured so that the
images have a superimposed field with each other. The images are
combined utilizing the superimposed field. The outside part of the
borderline of 180 degrees is the superimposed field. When two
lenses are employed for the image capturing apparatus, accordingly,
a lens has an angle of view equal to or larger than 180 degrees,
and preferably about 190 degrees so as to provide the superimposed
field. When four lenses are employed for the image capturing
apparatus, a lens has an angle of view of about 100 degrees so as
to provide an appropriate superimposed field.
[0043] With the structure employing four lenses illustrated on the
right in FIG. 16 and when images are captured with the image
capturing apparatus correctly being upright, the horizontal line is
captured just horizontally in the same manner described above. The
position of the vertical point recognized by the image capturing
person using the apparatus corresponds to the position of the
vertical point of the image capturing apparatus. The
above-described technique is widely known for combining captured
images to perform distortion correction on the combined image.
[0044] For example, when distortion correction is performed on a
plurality of images and the images are combined with each other in
the same manner as in Mercator projection, an image illustrated in
FIG. 18 is obtained. FIG. 18 is a diagram for explaining an image
obtained by performing distortion correction on images illustrated
in FIG. 17 and then combining the images with each other.
[0045] When the image capturing apparatus is placed so as to be
correctly upright and images are captured without the apparatus
being tilted, a correct image can be obtained having the straight
horizontal line as illustrated in FIG. 18 after distortion
correction and combination are simply performed on the captured
images. Specifically, images can be captured in the correct
vertical direction when securing the image capturing apparatus to a
certain retainer, and adjusting the horizontal and vertical lines
of the apparatus using a level, for example.
[0046] FIG. 19 is a diagram for explaining an image captured by the
multi-eye (two-eye) image capturing apparatus illustrated on the
right in FIG. 16, with the image capturing apparatus in a state of
being tilted. When a person holds the image capturing apparatus in
its hand(s), it is generally difficult to capture images
horizontally and vertically. FIG. 19 exemplifies images captured in
such a state. As illustrated in FIG. 19 the vertical points of the
images do not correspond to each other and the horizontal lines of
the images are distorted. As described above, if distortion
correction and combination are performed on images captured with
the image capturing apparatus in a state of being tilted without
consideration of the tilt, an image is obtained in which the
distortion exemplified in FIG. 19 is directly reflected, as
illustrated in FIG. 20.
[0047] If distortion correction and combination are performed on
images captured with the image capturing apparatus in a state of
being tilted without consideration of the tilt, the horizontal line
is curved like a graph of a trigonometric function as illustrated
in FIG. 20. In addition, the orthogonality of the image is broken.
To avoid these problems, correction should be performed in
consideration of a tilt angle in the vertical direction.
[0048] The principles of measurement of a tilt angle will now be
described with reference to FIG. 21. FIG. 21 is a schematic diagram
for explaining an acceleration sensor that measures a tilt
angle.
An acceleration sensor as illustrated in FIG. 21 is embedded in the
image capturing apparatus so as to measure how the image capturing
apparatus tilts in the vertical direction using a gravity
acceleration sensor.
[0049] FIG. 21 illustrates the outline of obtaining an angle using
a single-axis acceleration sensor. This is employed to simplify the
description. In FIG. 21, the sensor obtains only the tilt of the
image capturing apparatus within the plane including the central
axis of the lens surface in the two-eye structure. When a user
actually captures images, however, it is assumed that images may be
captured from the deviated angle from the plane as described above.
To solve this problem a three-axis acceleration sensor is employed
so as to measure the deviated angle from the plane including the
central plane of the lens surface. The image capturing apparatus
according to the embodiment will now be described in detail. FIG. 1
is a schematic block diagram for explaining the overall structure
of an electrical circuit of the image capturing apparatus according
to the embodiment of the present invention. As illustrated in FIG.
1, in this image capturing apparatus (hereinafter, also referred to
as a digital camera) 100, an image capturing element 1 (reference
numeral: 109), an image capturing element 2 (reference numeral:
110), a synchronous dynamic random access memory (SDRAM) 111, an
external storage device 112, and the acceleration sensor 113 are
coupled to a controller 10.
[0050] In the embodiment, two capturing elements (i.e., the two-eye
structure) are employed so as to obtain omni-directional images,
however, three or more capturing elements may be employed. When
three capturing elements are employed, a lens corresponding to the
image capturing element, which will be described with reference to
FIG. 2, does not require an angle of view equal to or larger than
180 degrees. Another angle of view of the lens may be employed as
appropriate. Wide-angle lenses including a fisheye lens are
typically employed for the lens. The image capturing apparatus is
not limited to the omni-directional image capturing apparatus.
Another image capturing apparatus capable of capturing images
covering 360 degrees in the horizontal direction may be used.
[0051] The controller 10 includes a central processing unit (CPU)
101, a static RANI (SRAM) 102, a read-only memory (ROM) 103, an
image processing block 104, an SDRAM interface (I/F) 105, an
external storage device I/F 106, and an external sensor I/F
107.
[0052] In the description of the embodiment, the image processing
block 104 performs typical image processing such as distortion
correction and pixel defect correction, and the CPU reads a
predetermined table or a computer program, thereby performing
correction processing in the vertical direction according to the
tilt of the digital camera 100. It is to be understood, however,
that the image processing block 104 may perform correction
processing in the vertical direction.
[0053] During capture, pieces of digitized image data are input to
the image processing block 104 of the controller 10 through the
image capturing element 1 (reference numeral: 109) and the image
capturing element 2 (reference numeral: 110). The pieces of image
data input are image-processed using the image processing block
104, the CPU 101, the SRAM 102, and the SDRAM 111, for example, and
finally stored in the external storage device 112. Examples of an
external storage device include CompactFlash (registered trademark)
or a Secure Digital (SD) memory.
[0054] In the controller 10, a universal serial bus (USB)
connection interface for connecting to an external device or a
wired or wireless network I/F for connecting to a network may be
provided. A conversion table, which will be described later, a
correction processing program for the conversion table, and a
processing program for calculation of vertical correction are
stored in the SRAM 102 or the SDRAM 111.
[0055] The acceleration sensor 113 is used for detecting a tilt of
the digital camera 100 during capture. This enables detection in
the tilting direction of the digital camera instantly and
readily.
[0056] The acceleration sensor 113 is a three-axis acceleration
sensor that detects acceleration of the digital camera 100 in the
three directions perpendicular to each other, i.e., the up and down
direction, the right and left direction, and the back and front
direction.
[0057] When a user holds the digital camera 100 in its hand(s) so
that the digital camera 100 remains still, the acceleration sensor
113 detects the gravity acceleration only.
[0058] When acceleration is detected only in the down direction in
the up and down direction, it is found that the up and down
direction of the digital camera 100 corresponds to the top and
bottom direction with respect to the ground. In other words, it is
known that the digital camera is held horizontally in the same
manner usually operating a digital camera.
[0059] When the image capturing apparatus tilts in the top and
bottom direction, the acceleration sensor 113 detects acceleration
in the right and left direction and the back and front direction
according to the actual tilting direction. A tilt angle of the
digital camera 100 can be obtained according to the magnitude of
acceleration in the up and down direction, the right and left
direction, and the back and front direction.
[0060] The omni-directional spherical image capturing apparatus
will now be described. FIG. 2 is an external side view of the
omni-directional spherical image capturing apparatus according to
the embodiment.
[0061] The present embodiment aims to generate the omni-directional
spherical image using the omni-directional spherical image
capturing apparatus capable of capturing images in all directions
from a capture point. That is to say, the omni-directional
spherical image capturing apparatus (digital camera) can capture
images in all directions viewed from the capture point.
[0062] The digital camera 100 (illustrated in FIG. 1) serving as an
omni-directional spherical image capturing apparatus, captures
images through the two capturing element 1 (reference numeral: 109)
and the image capturing element 2 (reference numeral: 110). The
respective capturing elements have a fisheye lens as an example of
wide-angle lenses having an angle of view exceeding 180 degrees.
The images captured through these two fisheye lenses have a
superimposed field with each other. A predetermined image
processing such as distortion correction is performed on the
images, and the resulting images are converted, and then combined
with each other, thereby generating an omni-directional spherical
image. Operations of the omni-directional spherical image capturing
apparatus according to the embodiment will now be described. FIG. 3
is a flowchart for explaining operations of the image capturing
apparatus according to the embodiment. An explanation will be given
on the operations of the omni-directional spherical image capturing
apparatus from the time when captured images are input to the time
when the images are stored in the external storage device 112 (FIG.
1) with reference to FIG. 3.
[0063] The acceleration sensor 113 detects a tilt angle of the
digital camera 100 at Step (hereinafter, noted as "S") 301. At
5302, the controller 10 reads out a conversion table stored in the
SDRAM 111, for example, according to a tilt angle of the digital
camera 100 detected at S301. The controller 10 then corrects the
conversion table in a predetermined correction method. The
correction method for the conversion table will be described
later.
[0064] At S303, two digitized fisheye images captured through the
image capturing element 1 (reference numeral: 109) and the image
capturing element 2 (reference numeral: 110) are input to the image
processing block 104. The image processing block 104 performs
typical image processing such as distortion correction. At S304,
the controller 10 converts the two captured fisheye images (like
the image as illustrated in FIG. 17) using the conversion table
corrected at S302. The conversion method will be described
later.
[0065] At S305, the controller 10 generates a composed
omni-directional spherical image utilizing a superimposed field
between the two images converted at S304. At S306, the controller
10 stores the omni-directional spherical image generated at S305 in
the external storage device 112 through the external storage device
I/F 106.
[0066] A projection relation of the fisheye lens will now be
described as an example of wide-angle lenses employed in the
digital camera according to the embodiment. FIG. 4 illustrates a
projection relation of a fisheye lens employed in the digital
camera according to the embodiment. FIG. 4(a) is an external side
view of the fisheye lens; FIG. 4(b) illustrates a projection
function fin a plan view of a captured image.
[0067] FIG. 17 illustrates an image captured through the fisheye
lens with an angle of view over 180 degrees, which is a captured
image of an almost hemispherical scene viewed from a capture
position. As illustrated in (a) and (b) of FIG. 4, the image is
generated with an image height h according to an angle of incidence
.theta.. The relation of the angle of incidence .theta. and the
image height h is defined by a projection function f
(h=f(.theta.)). The projection function varies depending on the
characteristic of the fisheye lens.
[0068] Examples of a technique (function) of projective
transformation include central projection, stereographic
projection, equidistant projection, equi-solid-angle projection,
and orthographic projection. The central projection is employed
when capturing images using a digital camera including a typical
angle of view. Other four methods above are employed in a digital
camera including a wide-angle lens with an ultra wide angle of view
such as a fisheye lens.
[0069] A format (form of equation) of the omni-directional
spherical image will now be described that is captured by the
digital camera according to the embodiment. FIG. 5 illustrates a
format of the omni-directional spherical image captured by the
digital camera according to the embodiment. FIG. 5(a) illustrates a
plane format; FIG. 5(b) illustrates a spherical format.
[0070] FIG. 5(a) illustrates a format for developing the
omni-directional spherical image into plane coordinates. As
illustrated in FIG. 5(a), the plane format is an image including
pixel values corresponding to angular coordinates with horizontal
angles from 0 to 360 degrees and vertical angles from 0 to 180
degrees. The angular coordinates are related to the points in
spherical coordinates as illustrated in FIG. 5(b), which resemble
latitude and longitude coordinates on a globe.
[0071] Plane coordinates of images captured through the fisheye
lens and spherical coordinates of the omni-directional spherical
image can be related to each other using the projection function f
(h=f(.theta.)) illustrated in FIG. 4. This enables conversion of
two images captured through the fisheye lens and combination
(composition) of the resulting images, thereby generating
omni-directional spherical images as illustrated in FIG. 5(a) and
(b).
[0072] The process of generating the omni-directional spherical
image will now be described with reference to an actual image
captured through the fisheye lens. FIG. 6 illustrates the outline
of generation processing of the omni-directional spherical image
captured by the digital camera according to the embodiment. FIGS.
6(a1) and (a2) are images captured with capturing elements through
the two fisheye lenses; FIGS. 6(b1) and (b2) are images converted
through a conversion table (corresponding to FIG. 5(a)); FIG. 6(c)
is an image generated by combining (compositing) the two converted
images (corresponding to FIG. 5(b)). The respective images
illustrated in FIGS. 6(a1) and (a2) captured with capturing
elements through the two fisheye lenses, which are also
schematically illustrated in FIG. 17, are converted into the images
illustrated in FIGS. 6(b1) and (b2). This conversion is performed
through the processing at S304 illustrated in FIG. 3, i.e., image
conversion processing in which the corrected conversion table is
used. At this point, the images illustrated in FIGS. 6(b 1) and
(b2) are presented in a corresponding manner to the
omni-directional spherical image format, that is, corresponding to
the images illustrated in FIG. 5(a).
[0073] Subsequently, the processing at S305 illustrated in FIG. 3
is performed. Specifically, two converted images are combined,
thereby generating the omni-directional spherical image. More
specifically, the two converted images are superimposed using a
superimposed field of the images illustrated in FIGS. 6(b1) and
(b2) as a key and composed, whereby an image illustrated in FIG.
6(c), i.e., the omni-directional spherical image is generated.
[0074] The correction method for the conversion table at S304
illustrated in FIG. 3 will now be described. FIG. 7 illustrates the
conversion table for the omni-directional spherical image captured
by the digital camera according to the embodiment.
[0075] FIG. 7(a) is a diagram for explaining a conversion table
representing a matrix of coordinates of a pre-conversion image and
a post-conversion image. FIG. 7(b) and (c) are diagrams for
explaining a correspondence relation of the coordinates of the
pre-conversion image (FIG. 7(b)) and the coordinates of
post-conversion image (FIG. 7(c)).
[0076] FIG. 7(a) illustrates the conversion table used for the
image processing at S304 illustrated in FIG. 3. The conversion
table includes data sets of coordinates of pixel values of the
post-conversion image (.theta., .phi.) and the corresponding
coordinates of pixel values of the pre-conversion image (x, y), for
every coordinates of the post-conversion image. The conversion
table here represents a tabular data structure, however, another
structure may be used as long as it is conversion data.
[0077] A post-conversion image is generated from the captured image
(pre-conversion image) according to the conversion table
illustrated in FIG. 7(a). Specifically, as illustrated in FIG. 7(b)
and (c), pixels in the post-conversion image are generated by
referring to the corresponding pixel values in the coordinates in
the pre-conversion image, based on the correspondence relation
between the pre-conversion image and the post-conversion image in
the conversion table (FIG. 7(a)).
[0078] An explanation will now be given on processing of correcting
the conversion table for the omni-directional spherical image
captured by the digital camera according to the embodiment,
according to the tilt of the digital camera. FIG. 8 is a flowchart
for explaining operations of correction processing on the
conversion table for the omni-directional spherical image captured
by the image capturing apparatus according to the embodiment.
[0079] At S801, the controller 10 obtains a camera tilt parameter
(.alpha., .beta.) according to a tilt value of the digital camera
100 detected by the acceleration sensor 113 (FIGS. 1). .alpha. and
.beta. in the camera tilt parameter (.alpha., .beta.) represent
rotation angles, which will be described later.
[0080] At S802, input values (.theta.1, .phi.1) of the conversion
table are set. In FIG. 8, values in a camera coordinate system are
presented as (.theta.0, .phi.0) and values in a global coordinate
system are presented as (.theta.1, .phi.1) so as to be
differentiated from (.theta., .phi.), which are parameter values
that depend on a coordinate system. Accordingly, values (.theta.1,
.phi.1) in the global coordinate system are set at S802.
[0081] At S803, the input values (.theta.1, .phi.1) in the global
coordinate system are transformed into the values (.theta.0,
.phi.0) in the camera coordinate system through calculation of
vertical correction by the controller 10. This calculation of
vertical correction will be described later.
[0082] At S804, converted values (.theta.0, .phi.0) in the camera
coordinate system are transformed into the coordinates (x, y) of
the pre-conversion image through the conversion table not yet
corrected (FIG. 7(a)). It should be noted that a conversion table
for generating a correct omni-directional spherical image needs to
be provided in advance. The correct omni-directional spherical
image is generated under the condition of the camera without being
tilted during capture. The conversion table for generating the
correct omni-directional spherical image needs to be stored in a
predetermined storage unit such as an SRAM or an SDRAM.
[0083] At S805, the controller 10 stores input values (.theta.1,
.phi.1) in the global coordinate system and coordinates (x, y) not
yet corrected, which are calculated in the end above, as a set of
coordinates corresponding to each other in the conversion table
already corrected.
[0084] At S806, the controller 10 determines whether any
outstanding input values (.theta.1, .phi.1) remain. In other words,
the controller 10 determines whether any outstanding input values
(.theta.1, .phi.1) in the global coordinate system, in which the
corresponding coordinates (x, y) not yet corrected is not
calculated remain. If the controller 10 determines that outstanding
input values (.theta.1, .phi.1) remain (Yes at S806), the
processing returns to S802 so that input values (.theta.1, .phi.1)
in the global coordinate system are set as a subsequent value.
[0085] If the controller 10 determines that no outstanding input
values (.theta.1, .phi.1) remain (No at S806), the processing ends.
In this case, the controller 10 completes calculation of
coordinates (x, y) not yet corrected corresponding to the pixels in
the format of the omni-directional spherical image that have input
values (.theta.1, .phi.1) in the global coordinate system as
coordinates.
[0086] The tilt of the digital camera 100 according to the
embodiment will now be described. FIG. 9 is a schematic diagram for
explaining the tilt of the image capturing apparatus according to
the embodiment.
[0087] In FIG. 9, the vertical direction corresponds to the z-axis
in a three-dimensional (x, y, z) Cartesian coordinates in the
global coordinate system. When this direction corresponds to the
vertical direction of the digital camera illustrated in FIG. 9, the
camera is in a state of not being tilted. When the direction does
not correspond to the vertical direction of the digital camera, the
digital camera is in a state of being tilted.
[0088] A tilt angle .alpha. in the gravity vector and a gradient
angle .beta. in the xy plane are obtained from the equation below
using an output of the acceleration sensor. In the equation, Ax
refers to a component value in the x0-axis direction in the camera
coordinate system of the acceleration sensor, Ay refers to a
component value in the y0-axis direction in the camera coordinate
system of the acceleration sensor, and Az refers to a component
value in the z0-axis direction in the camera coordinate system of
the acceleration sensor.
.alpha.=Arc tan (Ax/Ay)
.beta.=Arc cos (Az/ {square root over
(Ax.sup.2+Ay.sup.2+Az.sup.2))}
[0089] The calculation of vertical correction will now be
described. FIGS. 10(a) and 10(b) are diagrams for explaining an
example of calculation of vertical correction on the
omni-directional spherical image captured by the digital camera
according to the embodiment. FIG. 10(a) illustrates a camera
coordinate system; FIG. 10(b) illustrates a global coordinate
system.
[0090] In FIG. 10, three-dimensional Cartesian coordinates in the
global coordinate system are presented as (x1, y1, z1) and
spherical coordinates in the global coordinate system are presented
as (.theta.1, .phi.1). In addition, three-dimensional Cartesian
coordinates in the camera coordinate system are presented as (x0,
y0, z0) and spherical coordinates in the camera coordinate system
are presented as (.theta.0, .phi.0).
[0091] Spherical coordinates (.theta.1, .phi.1) are transformed
into spherical coordinates (.theta.0, .phi.0) through the equations
illustrated in FIG. 10. To correct a tilt, three-dimensional
Cartesian coordinates are used to perform rotational transformation
at first, thus spherical coordinates (.theta.1, .phi.1) are
transformed into three-dimensional Cartesian coordinates (x1, y1,
z1) through Equations (1) to (3) illustrated in FIG. 10.
[0092] Subsequently, a camera tilt parameter (.alpha., .beta.) is
used to transform coordinates (x1, y1, z1) in the global coordinate
system into coordinates (x0, y0, z0) in the camera coordinate
system through the rotational coordinate transformation illustrated
in FIG. 10 (Equation (4)). In other words, this equation (Equation
(4) illustrated in FIG. 10) is the definition of the camera tilt
parameter (.alpha., .beta.).
[0093] This means that the camera coordinate system is obtained by
rotating the global coordinate system by a around the z-axis, and
then rotating the global coordinate system by .beta. around the
x-axis. Finally, the three-dimensional Cartesian coordinates (x0,
y0, z0) in the camera coordinate system are transformed back to
spherical coordinates (.theta.0, .phi.0) in the camera coordinate
system through Equations (5) and (6) illustrated in FIG. 10.
[0094] Another example of calculation of vertical correction will
now be described. FIG. 11 illustrates another example of
calculation of vertical correction on the omni-directional
spherical image captured by the digital camera according to the
embodiment. FIG. 11(a) illustrates a camera coordinate system; FIG.
11(b) illustrates a global coordinate system.
[0095] In the present embodiment, the calculation of vertical
correction is accelerated. Equations (1) to (6) described above and
illustrated in FIG. 10 according to the embodiment may be presented
as Equations (7) to (14) illustrated in FIG. 11.
[0096] Specifically, the rotation a and rotation y around the
z-axis is the rotation .theta. itself in spherical coordinates
(.theta., .phi.). Calculation for rotational transformation can be
performed with simple addition or subtraction operations, without
transforming into Cartesian coordinate system, thereby accelerating
the calculation. Accordingly, transformation with Cartesian
coordinate system is required only for the rotational
transformation of the rotation .beta. around the x-axis. This
speeds up the calculation.
[0097] An explanation will now be given on another processing of
correcting the conversion table for the omni-directional spherical
image captured by the digital camera according to the embodiment,
according to the tilt of the digital camera. FIG. 12 is another
flowchart for explaining operations of correction processing
through the conversion table for the omni-directional spherical
image captured by the image capturing apparatus according to the
embodiment.
[0098] In the present embodiment, the correction processing on the
conversion table can be accelerated. In the embodiment illustrated
in FIG. 11, the correction processing on the conversion table is
performed while accelerating the calculation of vertical
correction. By contrast, the calculation of vertical correction is
omitted to accelerate the calculation in the present
embodiment.
[0099] With reference to FIG. 12, the correction processing on the
conversion table will be described. At S1201, a camera tilt
parameter (.alpha., .beta.) is obtained. This processing is the
same as the processing at S801 illustrated in FIG. 8.
[0100] At S1202, the conversion table corresponding to the value of
the camera tilt parameter (.alpha., .beta.) is obtained , then the
processing ends. Specifically, the calculation of vertical
correction is omitted by storing a plurality of conversion tables
so that the conversion tables provide different values depending on
the camera tilt parameter (.alpha., .beta.).
[0101] The camera tilt parameter (.alpha., .beta.) is a
three-dimensional real vector, in principle. If conversion tables
are provided only for specific camera tilt parameters (.alpha.,
.beta.), and the conversion table closest to the detected camera
tilt parameter (.alpha., .beta.) is used, all parameters are
covered with the provided tables. Alternatively, it is also
effective that a plurality of tables close to the detected camera
tilt parameter (.alpha., .beta.) are extracted and an interpolation
operation such as weighting or obtaining difference. This enables
correction of the conversion table only with an interpolation
operation, which is relatively simple, thereby suppressing the
processing on the operation.
[0102] The overall structure of the image capture system according
to another embodiment will now be described. FIG. 13 is a schematic
block diagram for explaining the overall structure of the image
capture system according to the embodiment of the present
invention. In the embodiment described above, the tilt correction
is performed in a spherical image capturing apparatus (digital
camera), however, the tilt correction is not limited to be
performed in a spherical image capturing apparatus.
[0103] As illustrated in FIG. 13, a wired or wireless communication
function is added on the spherical image capturing apparatus to
obtain an information processing apparatus capable of performing
image processing, such as a computer or a portable information
terminal (e.g., smartphone tablet personal computer). The images
before correction are transmitted to the information processing
apparatus and processing for correcting the tilt can be performed
in the destination information processing apparatus (destination
device).
[0104] The tilt correction processing, however, requires tilt
information on the spherical image capturing apparatus. The tilt
information on the spherical image capturing apparatus can be
detected only in the image capturing apparatus itself, thus the
tilt information needs to be transmitted along with the images
before correction. The conversion table may be stored in the
spherical image capturing apparatus and transmitted along with the
images. Alternatively, the conversion table may be transmitted to
the destination device in advance as preparation for capturing
(e.g., when the destination device recognizes the image capturing
apparatus, or when the image capturing apparatus recognizes the
destination device).
[0105] The conversion table does not need to be transmitted along
with the images every time. The conversion table may be transmitted
when the destination device requests the conversion table or checks
whether the conversion table is the latest one. In addition, the
conversion table may not be transmitted necessarily from the image
capturing apparatus. For example, the conversion table may be
uploaded to a website of the manufacturer that sells the image
capturing apparatus so as to be downloaded and stored in the
destination device. The conversion table transmitted from the image
capturing apparatus may be customized so as to suit the image
capturing apparatus.
[0106] An explanation will now be given on the destination device
200 in an image capture system according to another embodiment.
FIG. 14 is a schematic block diagram for explaining the overall
structure of an electronic circuit of the destination device 200 of
the image capture system according to another embodiment. In FIG.
14, an SDRAM 1411 and an external storage device 1412 are coupled
to a controller 140.
[0107] The controller 140 includes a CPU 1401, an SRAM 1402, a ROM
1403, an image processing block 1404, an SDRAM I/F 1405, and an
external storage device I/F 1406. The controller 140 also includes
a USB connection interface 1407 for connecting to an external
device and a wired or wireless network I/F 1408 for connecting to a
network.
[0108] In the description of the embodiment above, the image
processing block 1404 performs typical image processing such as
distortion correction and pixel defect correction, and the CPU 1401
reads a predetermined table or a computer program, thereby
performing correction processing in the vertical direction
according to the tilt of the digital camera 100. It is to be
understood, however, that the image processing block 1404 may
perform correction processing in the vertical direction.
[0109] Operations of the image capture system according to another
embodiment will now be described. FIG. 15 is a flowchart for
explaining operations of the image capture system according to
another embodiment. FIG. 15 illustrates that after a captured image
is input, the image is stored in a storage unit of the destination
device.
[0110] Processing is performed in the digital camera 100 firstly.
At S1501, a tilt angle of the digital camera 100 is detected by the
acceleration sensor 113 (FIG. 1). At S1502, two fisheye images (as
illustrated in FIG. 17) are transmitted to the destination device
200 such as a portable information terminal or a computer by the
controller 10. The fisheye images have been captured with the image
capturing element 1 (reference numeral: 109) and the image
capturing element 2 (reference numeral: 110) and digitized. The
transmission above is performed through the USB connection
interface 1407 or the network I/F 1408 as illustrated in FIG. 13.
Additionally, tilt information on the digital camera and the
conversion table are transmitted to the destination device 200
through the USB connection interface 1407 or the network I/F
1408.
[0111] As described above, the conversion table may be transmitted
in advance when the digital camera 100 and the destination device
200 recognize each other. The conversion table may be transmitted
to the destination device 200 only once, that is, the conversion
table does not need to be transmitted every time.
[0112] The conversion table is stored in the SDRAM 1411 coupled to
the digital camera 100, for example, and read out therefrom, then
transmitted. This is the end of processing performed in the digital
camera 100. The subsequent steps are performed in the destination
device 200.
[0113] At S1503, the conversion table is corrected in the
controller 140 through a predetermined correction method according
to the tilt information, which is the transmitted angular
information of the digital camera. The correction method for the
conversion table is the same in the embodiment described above. At
S1504, the transmitted fisheye images captured through the two
capturing elements are input to the image processing block 1404 of
the destination device 200.
[0114] In the image processing block 1404, typical image processing
such as distortion correction is performed. At S1505, the two
fisheye images are transformed by the controller 140 through the
conversion table corrected at S1503. The transformation method is
the same as in the embodiment described above.
[0115] At S1506, a composed omni-directional spherical image is
generated by the controller 140 utilizing the superimposed field of
the two images transformed at S1505. At S1507, the omni-directional
spherical image generated at S1506 is stored by the controller 140
in the external storage device 1412 through the external storage
device I/F 1406.
[0116] The operations relating to the flowcharts according to the
embodiments illustrated in FIGS. 3, 8, 12, and 15 can be executed
by a computer program. Specifically, the CPU (control circuit) 101
(FIG. 1) controlling operations of the image capturing apparatus or
the CPU (control circuit) 1401 (FIG. 14) controlling operations of
the destination device 200 loads various computer programs stored
in a recording medium such as the ROMs 103 and 1403, the SRAMs 102
and 1402, and sequentially executes the programs.
[0117] As described above, in the omni-directional spherical image
capturing apparatus, or in the image capture system, the vertical
direction is detected, and correction is performed on a conversion
table used for image processing, and an omni-directional spherical
image is generated through the corrected conversion table. This
eliminates the necessity of re-making the conversion table from the
beginning, thereby reducing the processing time.
[0118] Although the invention has been described with respect to
specific embodiments for a complete and clear disclosure, the
appended claims are not to be thus limited but are to be construed
as embodying all modifications and alternative constructions that
may occur to one skilled in the art that fairly fall within the
basic teaching herein set forth.
* * * * *